Re: Reduction Ratios

--- In CAD_CAM_EDM_DRO@yahoogroups.com, "roboticscnc" <gsm42@o...>
wrote:

>
>
> Hi Group,
> I am in the process of setting up the 24 volt servos on a 36" lathe.
> I'm using a Hitachi 2.1 amp, 500 pulse/rev encoder, 3/16" diameter
> shaft servo for the X axis. The Z axis servo is much larger with a 1/2
> inch shaft & 1000 pulses/rev encoder but with no other specs

available.

> I've heard that servos run at about 3:1 reduction but this doesn't
> allow for different pitch ballscrews. Can anyone tell me the distance
> the tool should move per revolution of the servo ? ( or how many
> revolutions/mm )
> Thanks,
> Glen

Glen. Here is the reaoning behind the 3:1 suggestions. At the normal
speeds most leadscrews need to be spun you need to furnish them with a
few hundred RPM. Granted that finer pitch ballscrews will act as part
of the speed reduction but most bigger equipment does not have 20 or
even 10 TPI screws. Typical is 5 or less. If you open a ballscrew
catalog only the really small ballscrews have fine pitches less than 7.

So lets take that as a starting point. With the typical 5 TPI ballscrew
You have to spin the motor times. Lets say you want to move the table
at 120IPM (2 inches per second). You would spin the motor 5 X 120 or
600 RPM. You probabaly won't cut at 120 IPM so lets say you cut at 45
IPM. That's only 225 RPM. Steppers operate just fine at lower RPM so
if you have enough torque you just direct couple them and go on.

Servos are different animals. Most have Rpm ranges from 1500 to 3000
PRM. Servos develop their best torque at high RPM. So if your target
is to run the servo in the middle of its RPM range during normal
cutting (where you need the most torque) with a 1500 RPM motor you need
to be at about 600 to 650 RPM for the 45IPM target. Note the ratio
between the actual leadscrew RPM and the target motor RPM (3:1). The
advantage is that not only do you get to run the servo in a better
portion of it's torque range you get a 3:1 torque amplification so you
can live with smaller servo motors. Servos wil turn at low rpm but
using typical encoder feedback they will start to get jerky (cog).

A not here about your encoder line counts. In a servo system the
encoder count determines the resolution of the system. The larger the
line count the better resolution you get and the better slow speed
characteritics you have BUT there is a problem. If you want get decent
linear speeds you have to provide lots more pulses per second to get
there. Here is an example using your 1000 line servos and running the
5 TPI leadscrew at 120 IPM:

There are 4000 counts from the 1000 line encoder. You there for need
4000 pulses to turn the motor one complete revolution. If it were
connected directly to the shaft you need 4000 * 5 or 20,000 pulses to
move the axis 1 inch. Since the common time value is seconds you need
20,000 pulse pers second to move 1 inch per second. That gives us 60
IPM. To hit our 120 spec we need to furnish 40,000 pulse per second;
right at the edge of a lot of PC based pulse systems. Now throw in the
3:1 ratio you need to run you motor in its range and you end up with
120,000 pulses per second.....

As you can see the high line count servo encoders create some
problems. This is where design tradeoffs occur. Do you change out the
encoders, direct couple the ballscrews, go with a step multiplier or
find a controller that can do 120,000 pulses per second? [ this space
left open for Fred to jump in]

Just be aware that designing a servo system is a set of actions and
interactions and changing any part of the equation will change the
answer!